In this example, the condition of 'no network connectivity' does not imply 'no network connectivity if things go wrong', it means 'no network connectivity'. So in his example, even when things are going *swimmingly*, the dictated standard of using an internet hosted monitoring infrastructure is not going to fly.
Standardisation is a term that can mean a lot or very little. If things were so simple that everyone who says 'monitoring' could be met with the same thing, the marketplace for that set of technologies would not be so varied. Now whether a single company can be met with one solution may be very possible depending on the company, but for other companies, it's not possible. For example, IBM has no such consistency within itself. Some financial institutions have very consistent desktop/laptop situation, but a highly varied datacenter picture (with mainframes, HFT, traditional back office stuff, etc). Retail chains are more usually in the neighborhood of high standardization being feasible. The problem comes in when people are appointed architects and fail to recognize the set of business needs appropriate for the business mission at hand (whether it's less regimented standards or conversely if the current state affairs should have been locked down but people were left without any support fending for themselves causing chaos). Standards don't have to be terrible and can save work, but you'll find a ton of folks disgruntled because frequently the standards are not selected well and compliance is not handled with any subtlety.
So what, you want me to know in detail the needs of every team in the organisation, and the capabilities of the 1300 pieces of software we already have deployed, and the opportunities presented by the 80,000 pieces of software we could potentially deploy?
The issue is not expecting you to know all of that. The problem comes in "In the meantime nobody's installing anything else if I get to hear about it", which suggests that you *don't* trust other people. Maybe you were oversimplifying for the sake of an internet forum, but the impression that someone coming in with a request to deviate is a 'fucker' suggests you have a pretty strong prejudice, and I've seen it frequently enough that folks are unreasonably held to a certain product or product line even as it makes zero sense for their use case. So I was frustrated to see a sentiment of 'I can't be bothered to know anything about the tool, but anyone who would not use the tool are fuckers' as I have had to deal with the fallout of precisely that sort of thinking. Giving the perception of inflexibility causes people not to change their plans so much as try not to call attention to how their plans are executing. I historically have found such instances and thought 'well, those calling the shots should be aware of the situation', inviting nothing but grief on me for pointing out a situation that I didn't even cause. Of course there are organizations that are a bit more loosely federated, and I find those to be pretty good places to get work done. Individual groups that are distinct from each other act with autonomy, as their work as recognized as meaningfully different than other groups. The problems occur in big multinationals with a horribly short set of standards to apply across the board across very different business functions. Note I tend to work in tech-heavy industries, so IT needs tend to be a bit more varied than some companies I have been where they really have no need for such stuff. (Note the premise of the article is that 'constant change' is good, only holds true for tech-heavy, for other companies IT needs to change are more like the need for the plumbing to be completely redone in the building... no reason to mess with it too much if not broken).
Conflict comes up when trying to paint everything with a wide brush. Non-tech SMBs are very different from Multinational tech companies, but we have articles and commentary trying to oversimplify that reality and make statements like 'change is always good' or 'constancy is always good', which sound good in context but the real world is more subtle than the perspective any specific commentator has.
the modern "everything is a file descriptor" is much closer to what Windows provides with its universal object handle metaphor.
I think this depends on your perspective.....
For my perspective it was 'everything is a file', meaning that some effort is expended to provide some sort of discoverable entry points into function. It's a matter of how it gets mapped to a namespace. This is an area where windows falls short as it does not do as much to model it's environment in an easily discoverable namespace. Of course things like 'ifconfig' and such violate the principle by needing to do magical invisible things for enumerating network devices for no particularly good reason.
Of course I hear you on the limitations of what happens in code once you have open references to 'things', and having to do some interesting unnatural things when the thing in question isn't modeled by a file handle and you want to deal with it or deal with it amongst a bunch of things that are modeled by a file handle.
Make sure every fucker else engages with that team to use that tool, and doesn't buy their own.
And this is why EAs are frowned upon and 'shadow' IT happens. The first team to come along evaluates candidate tools according to their needs,and are considered reasonable enough blokes. A second team that would *dare* not have the same needs as the first team are *presumed* fuckers until proven otherwise. Nevermind the first team's mission centered around managing a fleet of globally roaming Windows laptops, and the second team's mission focuses around densely packed rackmount servers in specific datacenters running linux, they should use *whatever* the first team came up with. Nevermind there is precisely zero overlap between the missions/staffing and there is *no* reason in that case for either team to follow the lead of the other. Note I speak from a position of particular soreness, as I have been involved in scenarios where a Windows-only vendor came first and declared standard. Then Linux came along and they were mandated to use some abandonware type thing from the vendor where they once compiled their codebase against libwine for the hell of it and continue selling it, even if it doesn't do anything, because they know they have a captive audience forced to license their stuff for any computer because of this precise phenomenon.
I don't need to know what the tool is, meet the vendor, take the training course... In the meantime nobody's installing anything else if I get to hear about it.
So you are *willfully* ignorant and don't want to understand, yet you won't let that stop you from having an overruling opinion over everyone else that may come along..... Damn, just damn.
Also, you are bending over your businesses to take it up the ass from the vendors. The vendors know when they are an exclusive provider and *will* control their pricing and their support accordingly. There's a reason why the very best support personnel at these companies are in the sales organization rather than the 'support' org. As soon as you are a done deal, they move on, leaving you with the general bottom of the barrel support and removing all those sweet discounts they offered when trying to compete for the business.
Yeah, from a hobby project it probably doesn't make much of a difference, but for commercial scenarios, I just don't see a good upside for either developers or MS.
Perhaps it's a sore spot because of a recent scenario where someone wrote a crappy.Net app and I'm having to do a different implementation because they weren't thinking about licensing and the licensing for Windows is damned near impossible for the scenario at hand. Also their code was crap, so it's not entirely terrible that their shortsightedness forced the issue.
Yeah, I think that goal may be misguided. For the same reason they probably live in terror of the reality that tons of XP/CE based platforms are alive in the world, they should be very afraid that Windows 10 IoT 'success' would look the same way: long defunct software responsible for critically important things that could give them a black eye at any moment. The upside? A relative pittance in licensing compared to their standard business model.
The things that are still XP/CE will forever be XP/CE. Windows 10 won't change that reality, just another generation of equipment that MS ends up hoping goes out of service ASAP if it does take off....
Embedded can be a very thankless market for software vendors. Anonymous software under the covers hidden well behind the brand of the hardware brand on the case just fits the market better for the most part.
If I write something while running Ubuntu, getting it to run in RHEL is no big deal. Changing to SuSE is no big deal. Changing to FreeBSD is *probably* not a big deal (as a developer). The distribution variability is more an issue for operators/administrators/users than anything on the developer side. Yes, if you develop against bleeding edge and then try to run it under an 'enterprise' lifecycle environment, the vintage of some of the offerings may frustrate, but that's life on any platform if you play in the deep end to start with. Just like with Windows, you do well to target established technologies rather than chase the latest and greatest, but it's still pretty workable either way. By the way, I have code that I wrote 15 years ago that still works on modern embedded platform. On the desktop, MS has handily shown that capability, but in embedded space, they have not had such a stellar track record for backwards compatibility.
I don't mean to say Pi is bad because it is an ARM platform, I'm saying better ARM platforms exist for the sorts of things Windows is equipped to deal with. Windows doesn't give full GPIO access to the thing anyway, and that's really the part of the Pi platform that distinguishes it from a horde of media stick type platforms with more powerful MediaTek SoCs.
'why' for developers and 'why' for microsoft as well.
For developers, MS is so mismatched to the sensibilities of the embedded space, business and technology wise. Picking up the ball and going home from one linux to another or even to something like a BSD is easy enough if you have to. If you commit to MS ecosystem, there's no where to go if things pan out poorly (e.g, Windows mobile, windows ce, windows phone (at least 7 was a dead end), Windows RT). MS has a terrible track record in this space, even when their wheelhouse of desktop application ecosystem has some relevance, where the Pi has pretty much no relevance (it may have video out, but there are better choices for even ARM based graphical systems than Pi). MS ecosystem is in general so *alien* compared to the rest of the industry, you *really* have to believe in it to commit. It's silly to bet your project on MS's technology and ongoing commitment to the platform in this market.
For MS, what do they hope to get out of this? They are coming into this from behind, against a competitor that gives away for free and where the entire ecosystem is tilted against them. They are going in to explore with no royalties, and no path to profit, or even revenue. Incidentally this has some resemblance to when they tried to break into 'supercomputing' nearly a decade ago, only to give up and let the resources mostly scatter to the winds when they figured out that there was no money to be made in the market, despite the prestige.
On the whole I greatly prefer versions that are year.month based so it is very easy to recognize how long in the tooth something is
Of course there are situations where that really does not matter and fretting about how old a piece of code is pointless worrying. For example, the code to decrypt a DVD hasn't change in an eternity, but there's no good reason to expect a change to be needed on that front. A library dedicated to that purpose might not change because there's no point, not because it's abandoned.
While there is a practical facet of that approach, it misses a lot of things.
For one, if you want *only* bugfixes and *nothing* to change functionally in some app you use (or your confidence in the upstream to make feature changes without regressions is low), then semantic versioning offers a promise that you can use to select what you pursue. libtool speaks strictly only to functional compatibility, but some want to track beyond that promise.
Similarly, it speaks specifically to how API/ABIs work, not things like expectations around it interacts with a human. For example, a UI redesign wouldn't necessarily mean anything according to libtool versioning system, but it may be a huge thing for people trying to use the software, and they may want a way to tell pre-redesign from post-redesign.
Now the duration for which one can chase such areas hoping for backports of fixes, that's probably a big way in which the promise of semantic versions fade quickly (at least without some sort of out-of-band data about how *long* the project in question promises to try to backport).
foo 1.2.3 and foo 1.3.4 might both release as independent stable branches. 1.2.3 might release after 1.3.4, and that would be ok. Ditto for foo 2.1.3 releasing before 1.9.20. So in terms of the upstream, it's perfectly capable of describing concurrent releases.
If you refer to giving room to linux distros to ascribe another tier of versioning, that's why packages append their own revision and flavor (e.g. -5.el6 and -10.el7 may refer to both different generations of the package targeted at RHEL/CentOS6 v. 7.
If you are instead referring to some sort of disagreement in how the project moves forward, that's a fork with a rename (e.g. Gnome v. MATE, Trinity v. KDE),
Not seeing how that argument stands against semantic versioning. Arguments can be made whether or not one can meaningfully honor the promise of the versioning for very large, complex projects, but at least for smaller projects it's a reasonable enough approach.
Note that here the concept is that controlling the flow of the crowd is the automation piece. Whether you enact it through 'traffic signals', alerts to mobile devices, or send the data to uniformed officers to tell them how to direct traffic, the core is the automation piece here. An individual human at a given point does not have the awareness of the bigger picture to make the correct decision on how to flow traffic. So the discussion of having a human versus a signal light or other mechanism is not pertinent to the discussion at hand.
But going back to ignoring signals, this is actually a case where they would likely be followed. There's no pressing reason to ignore the directions (if it is going well) and while a few might opt to ignore it if they were alone, we are talking about a very crowded scenario. Here if even if only 30% of people would honor the signs individually, the entire crowd would probably honor the signs seeing such a high volume of people doing so. If it were midnight and someone who would run it pulls up next to someone who is stopped, they'd probably at least wait a short while. Then there's the case where they pull up behind, then they are *really* unlikely to run the red (basically going off the road or something similarly abnormal), and that would probably be the situation in a crowded environment given even a relative few honoring the signals.
I have a DK2. I have a GTX 660 and an midrange i5 (ivy bridge). It's ok by me for the most part, though I would appreciate a bump. That said, Wii level graphics are still pretty cool to me, and that doesn't need anything beefier than a midrange setup today (not laptop GPUs mind you, but still). 75 fps is what the DK2 does and I'm probably in the population that won't even notice the bump to 90 fps, but I welcome it nonethelist.
The thing I was talking about was Abrash's wishlist specifications for the future of VR that some are taking to say 'oh, but until we have those currently impossible specs, VR isn't *really* going to happen, because our tech is still too poor'. People clamor for 'reality' but really entertainment comes well before perfect 'realism'.
The issue is 'who cares'. Just like we do gaming when we can't do photorealistic gaming today, we don't need absolute realism to have a compelling experience. The environment is immersive, even if not 'deceptively realistic', moreso than the status quo of rendering the same thing on a monitor/TV.
The ambition to be absolutely real, allowing a full experience is in my mind overrated. I don't want to walk or climb, I have plenty of opportunities to do that in real life (and take advantage of them). I however do thoroughly enjoy my DK2 as it is, just letting me look around. The ability to have the 'oculus touch' type of controls is appealing, but I personally do not have a lot of excitement over things like treadmills, spheres, etc. I of course would love some wind and acceleration applied to aid in the experience of things like roller coasters, driving, and flying, but really don't need full locomotion on a grand scale..
It seems like it's not a small percentage, but neither it is 'everyone'.
I do not, and never have had this issue. Never motion sick, never had simulator sickness. Have subjected to all sorts of things that make other folks feel uneasy if not outright puking, not felt a pang of discomfort.
Movies and games on TV do not look like we are looking through a window, yet it's still very nice.
The reason I say that is that the requirements are pretty steep, and getting too much into the requirements might be 'perfect being the enemy of the good'.
Therein lies the crux of the problem for Apple. The way in is basically to do a lot more work enabling concepts like group policies and also 'lighten up licenses' so that effectively people can get use of their work for less money. There isn't an obvious way forward for Apple.
They can hope that players will upend the industry for them in a way that aligns to their sensibilities, but bending their sensibilities to try to capture the way IT works as-is would be a losing proposition.
The original argument was saying enterprise was great because a single person represents 10,000 instead of those same 10,000 being represented by 10,000 people. The counter argument is that Apple excels precisely at getting consumers to decide on their platform in an individual fashion, so they have no reason to be attracted to such a prospect.
They did want to be in the enterprise and hence the XServe being created. They realized they just weren't aligned with the industry and the prospects were grim for return on investment for trying to change that. So they stopped doing things that required them to spend money when the returns may likely never happen.
However when Cisco and IBM want to fall all over themselves to 'partner' with Apple, Apple will take the free endorsement. Note that both the Cisco and IBM deals cost Apple approximately nothing, they just had to smile and nod and endorse it, and in exchange IBM and Cisco spend all the money/do all the work to enable iOS devices for their respective applications and even promise some of their salespeople will push the Apple story. There's no point in turning down those overtures, even if they won't work or have low chance of working, all the risk is carried by IBM and Cisco. Potential upside is Apple suddenly is a viable mainstream enterprise vendor, downside is that Cisco and/or IBM wasted their time and money, but Apple lost nothing.
So it's not so inconsistent. They'll gladly take money from enterprises, but they don't believe it's worth spending money to try at this juncture.
Both Apple and MS would profit greatly by a deal like this.
Nope, MS would only lose out. MS has business captive today, and doing what you describe would just weaken their stranglehold. Note that a great deal of what enterprises do with 'Active Directory' goes way beyond the stuff that non-MS platforms support when they integrate, and much of that other stuff does not trivially map to anything but MS's particular vision of describing capabilities. The capabilities may be there across the board, but they are just organized so differently, it would be some investment to try to be apples-to-apples in an unambiguous way.
If Apple could provide that "any problem, one number" experience, businesses would beat a path to their door.
Except that they wouldn't be that one number. It would be MS and Apple. On the OS, sure vendors provide front line support all the time. When you move up MS stack... You are going to be calling MS if you have a problem. Note that IBM is the only IT company to really have unambiguous success at the game you describe (e.g. POWER chips with AIX on top with DB2 and an unholy mess of stuff on top of it, or the mainframe ecosystem), and even then there's been significant signs of trouble there. For examples of other attempts in the industry, HP is a very notable example of trying *really* hard to get to that IBM story, but no sign of them getting anywhere near IBM's level. HP gets plenty of revenue in other ways, but not specifically in the all-in-one.
If Apple seriously wanted in the enterprise sector, businesses would flock to them, just as an alternative to what is out there.
And the problem is that they wouldn't. Businesses don't change unless they are forced to. Even then they want the change to be as slight as reasonably possible. The motivator for change is either unbelievably high risk with current environment or very well defined cost savings. There isn't any particularly strong sign of risk where Apple improves compared to status quo. For cost, passionate arguments can be had about TCO, but those are very subjective arguments that vary greatly circumstance to circumstance. In practice, businesses make decisions on concrete metrics like acquisition cost and recurring license fees. On that front, Apple doesn't have much room to be compelling and also have the margin to which they are accustomed.
Enterprise is a big uphill battle that really isn't as appealing as many would imagine. Support costs are sky high, clients have a great deal more leverage than individual consumers to drive negotiation off of 'list' pricing, and generally have decades of accumulated infrastructure and best practices to work inside. For vendors entrenched in the space or pro
All the money is free. That's the thing. Modern economies are made up in no small part of manipulating the sentiment of the participants. The money numbers are completely made up and pretty arbitrary. It's more a delicate game of manipulation than earnest tracking of resources. The concept of 'resources' is now so fluid that it's actually a really hard concept.
Money does grow on/in trees, at least paper money in part does.
But yes the sentiment that it would 'elevate everyone out of poverty' is quite probably wishful thinking. I could belive a sort of mass-psychology effect in play that can constrict productivity due to exaggerated sense of scarcity that could be surprisingly reversed in such a scenario, but I'd hope that is not the only sort of thing keeping us from caring and feeding for the populace.
Excel is at it's most sophisticated use in most cases an integrated pocket calculator. Yes Excel can do very potent things for people who know more, but most of the time I see it used as just a tabular text editor with sporadic people knowing how to make a cell auto-add other cells.
It is hubris to think that this particular skillset has more universal appeal than any other industry that was transformative. Just because it's important and everywhere, doesn't mean everyone is willing and able to become masters of it. Cars are also massively popular, and yet only a relative few are mechanics. Indoor plumbing is very popular, yet folks still call for plumbers when things go south. Popularity does not imply mastery.
Most people never use it after their academic life. They might implicitly use some of the common concepts without thinking which is good, but they'll not sit down to write out an algebra problem.
Slashdot Mirror
In this example, the condition of 'no network connectivity' does not imply 'no network connectivity if things go wrong', it means 'no network connectivity'. So in his example, even when things are going *swimmingly*, the dictated standard of using an internet hosted monitoring infrastructure is not going to fly.
Standardisation is a term that can mean a lot or very little. If things were so simple that everyone who says 'monitoring' could be met with the same thing, the marketplace for that set of technologies would not be so varied. Now whether a single company can be met with one solution may be very possible depending on the company, but for other companies, it's not possible. For example, IBM has no such consistency within itself. Some financial institutions have very consistent desktop/laptop situation, but a highly varied datacenter picture (with mainframes, HFT, traditional back office stuff, etc). Retail chains are more usually in the neighborhood of high standardization being feasible. The problem comes in when people are appointed architects and fail to recognize the set of business needs appropriate for the business mission at hand (whether it's less regimented standards or conversely if the current state affairs should have been locked down but people were left without any support fending for themselves causing chaos). Standards don't have to be terrible and can save work, but you'll find a ton of folks disgruntled because frequently the standards are not selected well and compliance is not handled with any subtlety.
So what, you want me to know in detail the needs of every team in the organisation, and the capabilities of the 1300 pieces of software we already have deployed, and the opportunities presented by the 80,000 pieces of software we could potentially deploy?
The issue is not expecting you to know all of that. The problem comes in "In the meantime nobody's installing anything else if I get to hear about it", which suggests that you *don't* trust other people. Maybe you were oversimplifying for the sake of an internet forum, but the impression that someone coming in with a request to deviate is a 'fucker' suggests you have a pretty strong prejudice, and I've seen it frequently enough that folks are unreasonably held to a certain product or product line even as it makes zero sense for their use case. So I was frustrated to see a sentiment of 'I can't be bothered to know anything about the tool, but anyone who would not use the tool are fuckers' as I have had to deal with the fallout of precisely that sort of thinking. Giving the perception of inflexibility causes people not to change their plans so much as try not to call attention to how their plans are executing. I historically have found such instances and thought 'well, those calling the shots should be aware of the situation', inviting nothing but grief on me for pointing out a situation that I didn't even cause. Of course there are organizations that are a bit more loosely federated, and I find those to be pretty good places to get work done. Individual groups that are distinct from each other act with autonomy, as their work as recognized as meaningfully different than other groups. The problems occur in big multinationals with a horribly short set of standards to apply across the board across very different business functions. Note I tend to work in tech-heavy industries, so IT needs tend to be a bit more varied than some companies I have been where they really have no need for such stuff. (Note the premise of the article is that 'constant change' is good, only holds true for tech-heavy, for other companies IT needs to change are more like the need for the plumbing to be completely redone in the building... no reason to mess with it too much if not broken).
Conflict comes up when trying to paint everything with a wide brush. Non-tech SMBs are very different from Multinational tech companies, but we have articles and commentary trying to oversimplify that reality and make statements like 'change is always good' or 'constancy is always good', which sound good in context but the real world is more subtle than the perspective any specific commentator has.
the modern "everything is a file descriptor" is much closer to what Windows provides with its universal object handle metaphor.
I think this depends on your perspective.....
For my perspective it was 'everything is a file', meaning that some effort is expended to provide some sort of discoverable entry points into function. It's a matter of how it gets mapped to a namespace. This is an area where windows falls short as it does not do as much to model it's environment in an easily discoverable namespace. Of course things like 'ifconfig' and such violate the principle by needing to do magical invisible things for enumerating network devices for no particularly good reason.
Of course I hear you on the limitations of what happens in code once you have open references to 'things', and having to do some interesting unnatural things when the thing in question isn't modeled by a file handle and you want to deal with it or deal with it amongst a bunch of things that are modeled by a file handle.
Make sure every fucker else engages with that team to use that tool, and doesn't buy their own.
And this is why EAs are frowned upon and 'shadow' IT happens. The first team to come along evaluates candidate tools according to their needs,and are considered reasonable enough blokes. A second team that would *dare* not have the same needs as the first team are *presumed* fuckers until proven otherwise. Nevermind the first team's mission centered around managing a fleet of globally roaming Windows laptops, and the second team's mission focuses around densely packed rackmount servers in specific datacenters running linux, they should use *whatever* the first team came up with. Nevermind there is precisely zero overlap between the missions/staffing and there is *no* reason in that case for either team to follow the lead of the other. Note I speak from a position of particular soreness, as I have been involved in scenarios where a Windows-only vendor came first and declared standard. Then Linux came along and they were mandated to use some abandonware type thing from the vendor where they once compiled their codebase against libwine for the hell of it and continue selling it, even if it doesn't do anything, because they know they have a captive audience forced to license their stuff for any computer because of this precise phenomenon.
I don't need to know what the tool is, meet the vendor, take the training course ... In the meantime nobody's installing anything else if I get to hear about it.
So you are *willfully* ignorant and don't want to understand, yet you won't let that stop you from having an overruling opinion over everyone else that may come along..... Damn, just damn.
Also, you are bending over your businesses to take it up the ass from the vendors. The vendors know when they are an exclusive provider and *will* control their pricing and their support accordingly. There's a reason why the very best support personnel at these companies are in the sales organization rather than the 'support' org. As soon as you are a done deal, they move on, leaving you with the general bottom of the barrel support and removing all those sweet discounts they offered when trying to compete for the business.
Yeah, from a hobby project it probably doesn't make much of a difference, but for commercial scenarios, I just don't see a good upside for either developers or MS.
Perhaps it's a sore spot because of a recent scenario where someone wrote a crappy .Net app and I'm having to do a different implementation because they weren't thinking about licensing and the licensing for Windows is damned near impossible for the scenario at hand. Also their code was crap, so it's not entirely terrible that their shortsightedness forced the issue.
Yeah, I think that goal may be misguided. For the same reason they probably live in terror of the reality that tons of XP/CE based platforms are alive in the world, they should be very afraid that Windows 10 IoT 'success' would look the same way: long defunct software responsible for critically important things that could give them a black eye at any moment. The upside? A relative pittance in licensing compared to their standard business model.
The things that are still XP/CE will forever be XP/CE. Windows 10 won't change that reality, just another generation of equipment that MS ends up hoping goes out of service ASAP if it does take off....
Embedded can be a very thankless market for software vendors. Anonymous software under the covers hidden well behind the brand of the hardware brand on the case just fits the market better for the most part.
If I write something while running Ubuntu, getting it to run in RHEL is no big deal. Changing to SuSE is no big deal. Changing to FreeBSD is *probably* not a big deal (as a developer). The distribution variability is more an issue for operators/administrators/users than anything on the developer side. Yes, if you develop against bleeding edge and then try to run it under an 'enterprise' lifecycle environment, the vintage of some of the offerings may frustrate, but that's life on any platform if you play in the deep end to start with. Just like with Windows, you do well to target established technologies rather than chase the latest and greatest, but it's still pretty workable either way. By the way, I have code that I wrote 15 years ago that still works on modern embedded platform. On the desktop, MS has handily shown that capability, but in embedded space, they have not had such a stellar track record for backwards compatibility.
I don't mean to say Pi is bad because it is an ARM platform, I'm saying better ARM platforms exist for the sorts of things Windows is equipped to deal with. Windows doesn't give full GPIO access to the thing anyway, and that's really the part of the Pi platform that distinguishes it from a horde of media stick type platforms with more powerful MediaTek SoCs.
'why' for developers and 'why' for microsoft as well.
For developers, MS is so mismatched to the sensibilities of the embedded space, business and technology wise. Picking up the ball and going home from one linux to another or even to something like a BSD is easy enough if you have to. If you commit to MS ecosystem, there's no where to go if things pan out poorly (e.g, Windows mobile, windows ce, windows phone (at least 7 was a dead end), Windows RT). MS has a terrible track record in this space, even when their wheelhouse of desktop application ecosystem has some relevance, where the Pi has pretty much no relevance (it may have video out, but there are better choices for even ARM based graphical systems than Pi). MS ecosystem is in general so *alien* compared to the rest of the industry, you *really* have to believe in it to commit. It's silly to bet your project on MS's technology and ongoing commitment to the platform in this market.
For MS, what do they hope to get out of this? They are coming into this from behind, against a competitor that gives away for free and where the entire ecosystem is tilted against them. They are going in to explore with no royalties, and no path to profit, or even revenue. Incidentally this has some resemblance to when they tried to break into 'supercomputing' nearly a decade ago, only to give up and let the resources mostly scatter to the winds when they figured out that there was no money to be made in the market, despite the prestige.
On the whole I greatly prefer versions that are year.month based so it is very easy to recognize how long in the tooth something is
Of course there are situations where that really does not matter and fretting about how old a piece of code is pointless worrying. For example, the code to decrypt a DVD hasn't change in an eternity, but there's no good reason to expect a change to be needed on that front. A library dedicated to that purpose might not change because there's no point, not because it's abandoned.
While there is a practical facet of that approach, it misses a lot of things.
For one, if you want *only* bugfixes and *nothing* to change functionally in some app you use (or your confidence in the upstream to make feature changes without regressions is low), then semantic versioning offers a promise that you can use to select what you pursue. libtool speaks strictly only to functional compatibility, but some want to track beyond that promise.
Similarly, it speaks specifically to how API/ABIs work, not things like expectations around it interacts with a human. For example, a UI redesign wouldn't necessarily mean anything according to libtool versioning system, but it may be a huge thing for people trying to use the software, and they may want a way to tell pre-redesign from post-redesign.
Now the duration for which one can chase such areas hoping for backports of fixes, that's probably a big way in which the promise of semantic versions fade quickly (at least without some sort of out-of-band data about how *long* the project in question promises to try to backport).
Got an example at hand?
foo 1.2.3 and foo 1.3.4 might both release as independent stable branches. 1.2.3 might release after 1.3.4, and that would be ok. Ditto for foo 2.1.3 releasing before 1.9.20. So in terms of the upstream, it's perfectly capable of describing concurrent releases.
If you refer to giving room to linux distros to ascribe another tier of versioning, that's why packages append their own revision and flavor (e.g. -5.el6 and -10.el7 may refer to both different generations of the package targeted at RHEL/CentOS6 v. 7.
If you are instead referring to some sort of disagreement in how the project moves forward, that's a fork with a rename (e.g. Gnome v. MATE, Trinity v. KDE),
Not seeing how that argument stands against semantic versioning. Arguments can be made whether or not one can meaningfully honor the promise of the versioning for very large, complex projects, but at least for smaller projects it's a reasonable enough approach.
Note that here the concept is that controlling the flow of the crowd is the automation piece. Whether you enact it through 'traffic signals', alerts to mobile devices, or send the data to uniformed officers to tell them how to direct traffic, the core is the automation piece here. An individual human at a given point does not have the awareness of the bigger picture to make the correct decision on how to flow traffic. So the discussion of having a human versus a signal light or other mechanism is not pertinent to the discussion at hand.
But going back to ignoring signals, this is actually a case where they would likely be followed. There's no pressing reason to ignore the directions (if it is going well) and while a few might opt to ignore it if they were alone, we are talking about a very crowded scenario. Here if even if only 30% of people would honor the signs individually, the entire crowd would probably honor the signs seeing such a high volume of people doing so. If it were midnight and someone who would run it pulls up next to someone who is stopped, they'd probably at least wait a short while. Then there's the case where they pull up behind, then they are *really* unlikely to run the red (basically going off the road or something similarly abnormal), and that would probably be the situation in a crowded environment given even a relative few honoring the signals.
I have a DK2. I have a GTX 660 and an midrange i5 (ivy bridge). It's ok by me for the most part, though I would appreciate a bump. That said, Wii level graphics are still pretty cool to me, and that doesn't need anything beefier than a midrange setup today (not laptop GPUs mind you, but still). 75 fps is what the DK2 does and I'm probably in the population that won't even notice the bump to 90 fps, but I welcome it nonethelist.
The thing I was talking about was Abrash's wishlist specifications for the future of VR that some are taking to say 'oh, but until we have those currently impossible specs, VR isn't *really* going to happen, because our tech is still too poor'. People clamor for 'reality' but really entertainment comes well before perfect 'realism'.
The issue is 'who cares'. Just like we do gaming when we can't do photorealistic gaming today, we don't need absolute realism to have a compelling experience. The environment is immersive, even if not 'deceptively realistic', moreso than the status quo of rendering the same thing on a monitor/TV.
The ambition to be absolutely real, allowing a full experience is in my mind overrated. I don't want to walk or climb, I have plenty of opportunities to do that in real life (and take advantage of them). I however do thoroughly enjoy my DK2 as it is, just letting me look around. The ability to have the 'oculus touch' type of controls is appealing, but I personally do not have a lot of excitement over things like treadmills, spheres, etc. I of course would love some wind and acceleration applied to aid in the experience of things like roller coasters, driving, and flying, but really don't need full locomotion on a grand scale..
It seems like it's not a small percentage, but neither it is 'everyone'.
I do not, and never have had this issue. Never motion sick, never had simulator sickness. Have subjected to all sorts of things that make other folks feel uneasy if not outright puking, not felt a pang of discomfort.
Movies and games on TV do not look like we are looking through a window, yet it's still very nice.
The reason I say that is that the requirements are pretty steep, and getting too much into the requirements might be 'perfect being the enemy of the good'.
"It's an App!" - Admiral Ackbar
Therein lies the crux of the problem for Apple. The way in is basically to do a lot more work enabling concepts like group policies and also 'lighten up licenses' so that effectively people can get use of their work for less money. There isn't an obvious way forward for Apple.
They can hope that players will upend the industry for them in a way that aligns to their sensibilities, but bending their sensibilities to try to capture the way IT works as-is would be a losing proposition.
The original argument was saying enterprise was great because a single person represents 10,000 instead of those same 10,000 being represented by 10,000 people. The counter argument is that Apple excels precisely at getting consumers to decide on their platform in an individual fashion, so they have no reason to be attracted to such a prospect.
They did want to be in the enterprise and hence the XServe being created. They realized they just weren't aligned with the industry and the prospects were grim for return on investment for trying to change that. So they stopped doing things that required them to spend money when the returns may likely never happen.
However when Cisco and IBM want to fall all over themselves to 'partner' with Apple, Apple will take the free endorsement. Note that both the Cisco and IBM deals cost Apple approximately nothing, they just had to smile and nod and endorse it, and in exchange IBM and Cisco spend all the money/do all the work to enable iOS devices for their respective applications and even promise some of their salespeople will push the Apple story. There's no point in turning down those overtures, even if they won't work or have low chance of working, all the risk is carried by IBM and Cisco. Potential upside is Apple suddenly is a viable mainstream enterprise vendor, downside is that Cisco and/or IBM wasted their time and money, but Apple lost nothing.
So it's not so inconsistent. They'll gladly take money from enterprises, but they don't believe it's worth spending money to try at this juncture.
Both Apple and MS would profit greatly by a deal like this.
Nope, MS would only lose out. MS has business captive today, and doing what you describe would just weaken their stranglehold. Note that a great deal of what enterprises do with 'Active Directory' goes way beyond the stuff that non-MS platforms support when they integrate, and much of that other stuff does not trivially map to anything but MS's particular vision of describing capabilities. The capabilities may be there across the board, but they are just organized so differently, it would be some investment to try to be apples-to-apples in an unambiguous way.
If Apple could provide that "any problem, one number" experience, businesses would beat a path to their door.
Except that they wouldn't be that one number. It would be MS and Apple. On the OS, sure vendors provide front line support all the time. When you move up MS stack... You are going to be calling MS if you have a problem. Note that IBM is the only IT company to really have unambiguous success at the game you describe (e.g. POWER chips with AIX on top with DB2 and an unholy mess of stuff on top of it, or the mainframe ecosystem), and even then there's been significant signs of trouble there. For examples of other attempts in the industry, HP is a very notable example of trying *really* hard to get to that IBM story, but no sign of them getting anywhere near IBM's level. HP gets plenty of revenue in other ways, but not specifically in the all-in-one.
If Apple seriously wanted in the enterprise sector, businesses would flock to them, just as an alternative to what is out there.
And the problem is that they wouldn't. Businesses don't change unless they are forced to. Even then they want the change to be as slight as reasonably possible. The motivator for change is either unbelievably high risk with current environment or very well defined cost savings. There isn't any particularly strong sign of risk where Apple improves compared to status quo. For cost, passionate arguments can be had about TCO, but those are very subjective arguments that vary greatly circumstance to circumstance. In practice, businesses make decisions on concrete metrics like acquisition cost and recurring license fees. On that front, Apple doesn't have much room to be compelling and also have the margin to which they are accustomed.
Enterprise is a big uphill battle that really isn't as appealing as many would imagine. Support costs are sky high, clients have a great deal more leverage than individual consumers to drive negotiation off of 'list' pricing, and generally have decades of accumulated infrastructure and best practices to work inside. For vendors entrenched in the space or pro
All the money is free. That's the thing. Modern economies are made up in no small part of manipulating the sentiment of the participants. The money numbers are completely made up and pretty arbitrary. It's more a delicate game of manipulation than earnest tracking of resources. The concept of 'resources' is now so fluid that it's actually a really hard concept.
Money does grow on/in trees, at least paper money in part does.
But yes the sentiment that it would 'elevate everyone out of poverty' is quite probably wishful thinking. I could belive a sort of mass-psychology effect in play that can constrict productivity due to exaggerated sense of scarcity that could be surprisingly reversed in such a scenario, but I'd hope that is not the only sort of thing keeping us from caring and feeding for the populace.
Excel is at it's most sophisticated use in most cases an integrated pocket calculator. Yes Excel can do very potent things for people who know more, but most of the time I see it used as just a tabular text editor with sporadic people knowing how to make a cell auto-add other cells.
It is hubris to think that this particular skillset has more universal appeal than any other industry that was transformative. Just because it's important and everywhere, doesn't mean everyone is willing and able to become masters of it. Cars are also massively popular, and yet only a relative few are mechanics. Indoor plumbing is very popular, yet folks still call for plumbers when things go south. Popularity does not imply mastery.
People are forced to do algebra in high school.
Most people never use it after their academic life. They might implicitly use some of the common concepts without thinking which is good, but they'll not sit down to write out an algebra problem.